Skip to content

Latest commit

 

History

History
107 lines (92 loc) · 4.71 KB

1-setup.md

File metadata and controls

107 lines (92 loc) · 4.71 KB
Raw

1. Setting Up the Development Environment for AiidaDFTK Calculations

In this tutorial, we will guide you through setting up a development environment for running DFTK calculations using AiiDA. The tutorial is divided into two main sections:

  1. Julia Side: Creating a Julia project and installing AiidaDFTK.
  2. Python Side: Setting up aiida-core, installing the aiida-DFTK plugin, defining an AiiDA computer, code, and managing pseudopotentials.

Part 1: Julia Side

Preliminary: Installing Julia

In this tutorial we will call the computer where you want to run the DFTK calculations jed. On jed first install Julia. For this download the Julia binaries and follow the installation instructions. In this tutorial we will assume Julia is available under the path path/to/julia.

Steps to Set Up a Julia Project

  1. Launch a Julia REPL: E.g. by executing path/to/julia.
  2. Activate a New Environment: In the Julia REPL, create a new environment (essentially a new Julia project) using the following commands: 
    import Pkg
Pkg.activate("path/to/new/environment")
    Replace path/to/new/environment with the directory where you wish to create the new environment. This will generate a folder containing Project.toml and Manifest.toml files.
  3. Install AiidaDFTK: While still in the Julia REPL, initialize the project by adding packages you'll need. For this tutorial we will only need AiidaDFTK. 
    import Pkg
Pkg.add("AiidaDFTK")
  4. Optional configuration: See the details in the DFTK documentation on running DFTK on clusters on how to set up a LocalPreferences.toml file. This file allows to specify a cluster-specific configuration for libraries such as BLAS, LAPACK, FFT etc. This is optional, but can improve performance.

The new Julia project is now ready and includes AiidaDFTK and its dependency DFTK. You can now proceed with DFTK calculations in this isolated environment.

Part 2: Python Side

Step 1: Setting Up aiida-core

Follow the official aiida-core setup tutorial. Make sure you have a working aiida-core installation.

Step 2: Installing the aiida-DFTK Plugin

Install the aiida-dftk plugin repository from github and install it locally:

pip install git+https://github.com/aiidaplugins/aiida-dftk.git

Step 3: Defining an AiiDA Computer

For a detailed guide on defining a new computer in AiiDA, consult the official AiiDA tutorial. In this tutorial we will assume the new computer is available under the name jed.

Step 4: Register DFTK as an Aiida code

DFTK, when run through AiidaDFTK, should be registered as a code in the AiiDA database. This can be done using:

verdi code create core.code.installed \
    --label DFTK \
    --description "Julia DFTK" \
    --computer jed \
    --filepath-executable "path/to/julia" \
    --prepend-text 'export JULIA_PROJECT="path/to/new/environment"' \
    --default-calc-job-plugin dftk \
    --append-text '' \
    --no-use-double-quotes

The first two flags are description, the third specifies the computer from Step 3, the fourth and fifth the path to the Julia executable from Part 1 as well as the setup (prepended commands), which needs to be done to select the correct project folder with AiidaDFTK from Part 1. In this case this consists simply of exporting the environment variable, which tells Julia, which project folder to use. In some setup additional steps might be needed here, such as loading modules, which provide Julia or one of the vendor-specific libraries configured in Part 1, Step 4. The last three flags select dftk as the default input plugin and specify to use the default values for append_text and use_double_quotes.

Step 5: Managing Pseudopotentials

DFTK supports both HGH and UPF pseudopotentials. For realistic calculations, UPF format is recommended. We suggest using pseudo-dojo, which can be installed using the aiida-pseudo plugin. Follow this detailed tutorial for more information. Here's how to install pseudo-dojo 0.5 PBE pseudopotentials with scalar relativistic effects in UPF format:

aiida-pseudo install pseudo-dojo -v 0.5 -x PBE -r SR -f upf

Congratulations! You've successfully set up your development environment for AiidaDFTK and are ready to proceed with your calculations.